QuickDraw Time

metafile <- './quickdraw classification.csv'
meta <- read_csv(metafile, skip = 1)

Parsed with column specification:
cols(
  name = col_character(),
  animal = col_logical(),
  food = col_logical(),
  fruit = col_logical(),
  veggie = col_logical(),
  building = col_logical(),
  vechicle = col_logical(),
  bug = col_logical(),
  bird = col_logical(),
  plant = col_logical()
)

name_parts <- str_split_fixed(meta$name, coll("."), 2)
meta$basename <- name_parts[, 1]

bugs <- meta %>% filter(bug == TRUE)

Warning message:
In scan(file = file, what = what, sep = sep, quote = quote, dec = dec,  :
  EOF within quoted string

filename <- 'data/dog.stats.csv'
dogs_org <- read_csv(filename)

Parsed with column specification:
cols(
  key_id = col_double(),
  recognized = col_character(),
  word = col_character(),
  stroke_count = col_integer(),
  countrycode = col_character(),
  drawing_time = col_integer(),
  drawing_time_pause = col_integer(),
  drawing_time_draw = col_integer(),
  stroke_in_order = col_integer(),
  drawing_time_min = col_integer(),
  drawing_time_max = col_integer()
)

filename <- 'data/cat.stats.csv'
cats_org <- read_csv(filename)

Parsed with column specification:
cols(
  key_id = col_double(),
  recognized = col_character(),
  word = col_character(),
  stroke_count = col_integer(),
  countrycode = col_character(),
  drawing_time = col_integer(),
  drawing_time_pause = col_integer(),
  drawing_time_draw = col_integer(),
  stroke_in_order = col_integer(),
  drawing_time_min = col_integer(),
  drawing_time_max = col_integer()
)

Clean Data

prepare_data <- function (data) {
  new_data <- data %>% mutate(drawing_time_seconds = drawing_time / 1000, drawing_time_draw_seconds = drawing_time_draw / 1000, drawing_time_pause_seconds = drawing_time_pause / 1000)
  
  new_data <- new_data %>% 
    # a few have time values not in incremental
    filter(stroke_in_order == 0) %>% 
    # a few have weird times
    filter(drawing_time_min >= 0) %>% 
    filter(drawing_time_max > 0) %>%
    filter(drawing_time_draw > 100)
  return(new_data)
}

dogs <- prepare_data(dogs_org)
cats <- prepare_data(cats_org)

dogs %>% filter(recognized == 'True') %>%  filter(drawing_time_seconds < 25) %>%  #filter(stroke_count < 4) %>%
  ggplot(aes(x = drawing_time_seconds, y = (..count..)/sum(..count..))) +
  geom_histogram(binwidth = 0.1, fill = "#607d8b") +
  scale_y_continuous(labels = scales::percent) +
  scale_x_continuous(limits = c(0,20)) +
  labs(title = "Dog Total Drawing Time", y = "", x = "") +
  theme_joy()
ggsave("dog_total.png", width = 6, height = 5)

rr dogs_cats <- rbind(dogs, cats)

rr dogs_cats %>% filter(drawing_time_draw_seconds <= 20) %>% ggplot(aes(x = drawing_time_draw_seconds, y = word))+ geom_joy(scale = 4) + theme_joy() + scale_y_discrete(expand = c(0.01, 0)) + scale_x_continuous(expand = c(0, 0))

rr dogs_cats_rec_sum <- dogs_cats %>% group_by(word, recognized) %>% summarise(count = n()) %>% mutate(freq = count / sum(count))

rr cats_rec <- cats %>% filter(recognized == ‘True’) dogs_rec <- dogs %>% filter(recognized == ‘True’)

rr dogs_cats_rec <- rbind(dogs_rec, cats_rec)

rr dogs_cats_rec %>% ggplot(aes(x = word, y = drawing_time_pause_seconds)) + geom_boxplot()

rr summary(dogs_rec$drawing_time_pause_seconds)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  0.000   1.845   3.154   3.418   4.705 133.600 

rr summary(cats_rec$drawing_time_pause_seconds)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  0.000   2.561   3.699   4.118   5.291 261.400 

rr

#dogs_cats <- rbind(cat_drawings_rec, dog_drawings_rec)

rr NA

rr dogs_cats %>% filter(drawing_time_seconds > 20) %>% count() r dogs_cats %>% filter(drawing_time_seconds > 20) %>% count() / nrow(dogs_cats)

rr summary(dogs_rec$drawing_time_seconds)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  0.107   8.097  10.460  10.830  13.200 143.900 

rr summary(cats_rec$drawing_time_seconds)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  0.114   6.744   8.963   9.593  12.120 267.000 

rr dogs_cats %>% filter(drawing_time_seconds > 20) %>% ggplot(aes(x = word, y = drawing_time_pause_seconds)) + geom_boxplot()

rr NA

rr pg

$data
$data[[1]]

$data[[2]]

$layout
<ggproto object: Class Layout>
    facet: <ggproto object: Class FacetNull, Facet>
        compute_layout: function
        draw_back: function
        draw_front: function
        draw_labels: function
        draw_panels: function
        finish_data: function
        init_scales: function
        map: function
        map_data: function
        params: list
        render_back: function
        render_front: function
        render_panels: function
        setup_data: function
        setup_params: function
        shrink: TRUE
        train: function
        train_positions: function
        train_scales: function
        vars: function
        super:  <ggproto object: Class FacetNull, Facet>
    finish_data: function
    get_scales: function
    map: function
    map_position: function
    panel_layout: data.frame
    panel_ranges: list
    panel_scales: list
    render: function
    render_labels: function
    reset_scales: function
    setup: function
    train_position: function
    train_ranges: function
    xlabel: function
    ylabel: function
    super:  <ggproto object: Class Layout>

$plot

rr dogs_rec <- dogs_rec %>% mutate(time_bin = floor(drawing_time_draw_seconds)) dog_bins <- dogs_rec %>% group_by(time_bin) %>% summarise(count = n()) %>% mutate( freq = count / sum(count))

rr cats_rec <- cats_rec %>% mutate(time_bin = floor(drawing_time_draw_seconds)) cat_bins <- cats_rec %>% group_by(time_bin) %>% summarise(count = n()) %>% mutate( freq = count / sum(count))

rr dog_bins %>% ggplot() + geom_bar(data = dog_bins, mapping = aes(x = time_bin, y = freq), stat=‘identity’, fill = ‘blue’, alpha = 0.6) + geom_bar(data = cat_bins, mapping = aes(x = time_bin, y = freq), stat=‘identity’, fill = ‘red’, alpha = 0.6) + scale_x_continuous(limits=c(0, 25))

Sample Data

Test if the different n sizes affecting data.

rr cats_rec_sample <- cats_rec %>% sample_n(50000) dogs_rec_sample <- dogs_rec %>% sample_n(50000)

Pause Time

rr mean(cats_rec$drawing_time_seconds)

[1] 9.593239

rr mean(dogs_rec$drawing_time_seconds)

[1] 10.82946

rr mean(cats_rec$drawing_time_draw_seconds)

[1] 5.474867

rr mean(dogs_rec$drawing_time_draw_seconds)

[1] 7.411101

rr mean(cats_rec$stroke_count)

[1] 9.522221

rr mean(dogs_rec$stroke_count)

[1] 7.030852

rr dogs_rec %>% filter(drawing_time_draw_seconds < 30) %>% ggplot(aes(x = drawing_time_draw_seconds, y = stroke_count)) + geom_point(alpha = 1 / 100)

rr NA

rr cats_rec %>% ggplot(aes(x = drawing_time_draw_seconds, y = stroke_count)) + geom_point(alpha = 1 / 10)

rr NA

rr cats_rec %>% filter(drawing_time_pause_seconds < 20) %>% ggplot(aes(x = drawing_time_pause_seconds, y = stroke_count)) + geom_point(alpha = 1 / 10) + geom_smooth(method=‘lm’)

rr cats_rec_filtered <- cats_rec %>% filter(drawing_time_pause_seconds < 20)

rr fit <- lm(drawing_time_pause_seconds ~ stroke_count , data = cats_rec_filtered) summary(fit)

Call:
lm(formula = drawing_time_pause_seconds ~ stroke_count, data = cats_rec_filtered)

Residuals:
     Min       1Q   Median       3Q      Max 
-18.5908  -1.0138  -0.2461   0.8020  13.8677 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept)  0.301497   0.014292   21.09   <2e-16 ***
stroke_count 0.400271   0.001411  283.67   <2e-16 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 1.563 on 102933 degrees of freedom
Multiple R-squared:  0.4388,    Adjusted R-squared:  0.4387 
F-statistic: 8.047e+04 on 1 and 102933 DF,  p-value: < 2.2e-16

rr countrycodes_dogs <- dogs_rec %>% group_by(countrycode) %>% summarise(country_count = n()) %>% arrange(-country_count) countrycodes_cats <- cats_rec %>% group_by(countrycode) %>% summarise(country_count = n()) %>% arrange(-country_count)

rr top_countrycodes <- countrycodes_dogs %>% head(n = 30)

rr dogs_rec_top_countries %>% filter(drawing_time_draw_seconds < 30) %>% ggplot(aes(x = drawing_time_draw_seconds)) + facet_wrap(~ countrycode) + geom_histogram(aes(y=(..count..)/tapply(..count..,..PANEL..,sum)[..PANEL..]), binwidth = 1)

rr dogs_rec_top_countries %>% filter(drawing_time_draw_seconds < 30) %>% ggplot(aes(x = drawing_time_draw_seconds)) + facet_wrap(~ countrycode) + geom_histogram(aes(y=(..count..)/tapply(..count..,..PANEL..,sum)[..PANEL..]), binwidth = 1)

rr NA

rr dogs_rec_top_countries %>% filter(drawing_time_pause_seconds < 15) %>% ggplot(aes(x = drawing_time_pause_seconds)) + facet_wrap(~ countrycode) + geom_histogram(aes(y=(..count..)/tapply(..count..,..PANEL..,sum)[..PANEL..]), binwidth = 1)

rr NA

Category sub-sections

read_names <- function (names) {
  data_path <- './data/'
  all_cates <- vector("list", nrow(names))  
  
  for(i in 1:nrow(names)) {
    name_file <- names[i, "basename"]
    full_path <- paste(data_path, name_file, ".stats.csv", sep='')
    name <- read_csv(full_path)
    all_cates[[i]] <- name
  }
  
  all_cates_df <- do.call("rbind", all_cates)
  
  all_cates_df_clean <- prepare_data(all_cates_df)

  all_cates_df_clean_rec <- all_cates_df_clean %>% filter(recognized == 'True')
  
  return(all_cates_df_clean_rec)
  
}

all_bugs_df <- read_names(bugs)

Parsed with column specification:
cols(
  key_id = col_double(),
  recognized = col_character(),
  word = col_character(),
  stroke_count = col_integer(),
  countrycode = col_character(),
  drawing_time = col_integer(),
  drawing_time_pause = col_integer(),
  drawing_time_draw = col_integer(),
  stroke_in_order = col_integer(),
  drawing_time_min = col_integer(),
  drawing_time_max = col_integer()
)
Parsed with column specification:
cols(
  key_id = col_double(),
  recognized = col_character(),
  word = col_character(),
  stroke_count = col_integer(),
  countrycode = col_character(),
  drawing_time = col_integer(),
  drawing_time_pause = col_integer(),
  drawing_time_draw = col_integer(),
  stroke_in_order = col_integer(),
  drawing_time_min = col_integer(),
  drawing_time_max = col_integer()
)
Parsed with column specification:
cols(
  key_id = col_double(),
  recognized = col_character(),
  word = col_character(),
  stroke_count = col_integer(),
  countrycode = col_character(),
  drawing_time = col_integer(),
  drawing_time_pause = col_integer(),
  drawing_time_draw = col_integer(),
  stroke_in_order = col_integer(),
  drawing_time_min = col_integer(),
  drawing_time_max = col_integer()
)
Parsed with column specification:
cols(
  key_id = col_double(),
  recognized = col_character(),
  word = col_character(),
  stroke_count = col_integer(),
  countrycode = col_character(),
  drawing_time = col_integer(),
  drawing_time_pause = col_integer(),
  drawing_time_draw = col_integer(),
  stroke_in_order = col_integer(),
  drawing_time_min = col_integer(),
  drawing_time_max = col_integer()
)
Parsed with column specification:
cols(
  key_id = col_double(),
  recognized = col_character(),
  word = col_character(),
  stroke_count = col_integer(),
  countrycode = col_character(),
  drawing_time = col_integer(),
  drawing_time_pause = col_integer(),
  drawing_time_draw = col_integer(),
  stroke_in_order = col_integer(),
  drawing_time_min = col_integer(),
  drawing_time_max = col_integer()
)
Parsed with column specification:
cols(
  key_id = col_double(),
  recognized = col_character(),
  word = col_character(),
  stroke_count = col_integer(),
  countrycode = col_character(),
  drawing_time = col_integer(),
  drawing_time_pause = col_integer(),
  drawing_time_draw = col_integer(),
  stroke_in_order = col_integer(),
  drawing_time_min = col_integer(),
  drawing_time_max = col_integer()
)
Parsed with column specification:
cols(
  key_id = col_double(),
  recognized = col_character(),
  word = col_character(),
  stroke_count = col_integer(),
  countrycode = col_character(),
  drawing_time = col_integer(),
  drawing_time_pause = col_integer(),
  drawing_time_draw = col_integer(),
  stroke_in_order = col_integer(),
  drawing_time_min = col_integer(),
  drawing_time_max = col_integer()
)

bugs_agg <- all_bugs_df %>% group_by(word) %>% summarise(count = n(),
            stroke_count_mean = mean(stroke_count),
            stroke_count_med = median(stroke_count),
            drawing_time_pause_sec_mean = mean(drawing_time_pause_seconds), 
            drawing_time_pause_sec_med = median(drawing_time_pause_seconds), 
            drawing_time_draw_sec_mean = mean(drawing_time_draw_seconds), 
            drawing_time_draw_sec_med = median(drawing_time_draw_seconds)) %>%
  arrange(-drawing_time_draw_sec_mean)

all_bugs_df %>% 
ggplot(aes(x = drawing_time_draw_seconds, y = factor(word, levels = bugs_agg$word), fill = factor(word, levels = bugs_agg$word))) +
geom_joy(size = 0.55 ) +
  theme_joy() +
  scale_x_continuous(limits=c(1, 20), expand = c(0.01, 0)) +
  scale_y_discrete(expand = c(0.01, 0)) + 
  scale_fill_brewer(guide = FALSE, palette = "Greys", direction = -1) + 
  labs(title = 'Bugs', x = '', y = '')

ggsave('bugs.png', width = 8, height = 12)

fruits <- meta %>% filter(fruit == TRUE)
all_fruits <- read_names(fruits)

|=========                                                                      |  11%           
|=========                                                                      |  12%           
|==========                                                                     |  12%           
|==========                                                                     |  13%    1 MB
|===========                                                                    |  14%    1 MB
|===========                                                                    |  14%    1 MB
|============                                                                   |  15%    1 MB
|=============                                                                  |  16%    1 MB
|=============                                                                  |  17%    1 MB
|==============                                                                 |  17%    1 MB
|==============                                                                 |  18%    1 MB
|===============                                                                |  19%    1 MB
|===============                                                                |  19%    1 MB
|================                                                               |  20%    1 MB
|=================                                                              |  21%    1 MB
|=================                                                              |  22%    1 MB
|==================                                                             |  22%    1 MB
|==================                                                             |  23%    1 MB
|===================                                                            |  24%    1 MB
|===================                                                            |  24%    1 MB
|====================                                                           |  25%    1 MB
|=====================                                                          |  26%    1 MB
|=====================                                                          |  27%    2 MB
|======================                                                         |  27%    2 MB
|======================                                                         |  28%    2 MB
|=======================                                                        |  29%    2 MB
|=======================                                                        |  29%    2 MB
|========================                                                       |  30%    2 MB
|=========================                                                      |  31%    2 MB
|=========================                                                      |  31%    2 MB
|==========================                                                     |  32%    2 MB
|==========================                                                     |  33%    2 MB
|===========================                                                    |  34%    2 MB
|===========================                                                    |  34%    2 MB
|============================                                                   |  35%    2 MB
|=============================                                                  |  36%    2 MB
|=============================                                                  |  36%    2 MB
|==============================                                                 |  37%    2 MB
|==============================                                                 |  38%    2 MB
|===============================                                                |  39%    2 MB
|===============================                                                |  39%    3 MB
|================================                                               |  40%    3 MB
|================================                                               |  41%    3 MB
|=================================                                              |  41%    3 MB
|==================================                                             |  42%    3 MB
|==================================                                             |  43%    3 MB
|===================================                                            |  44%    3 MB
|===================================                                            |  44%    3 MB
|====================================                                           |  45%    3 MB
|====================================                                           |  46%    3 MB
|=====================================                                          |  46%    3 MB
|======================================                                         |  47%    3 MB
|======================================                                         |  48%    3 MB
|=======================================                                        |  49%    3 MB
|=======================================                                        |  49%    3 MB
|========================================                                       |  50%    3 MB
|========================================                                       |  51%    3 MB
|=========================================                                      |  51%    3 MB
|==========================================                                     |  52%    3 MB
|==========================================                                     |  53%    4 MB
|===========================================                                    |  54%    4 MB
|===========================================                                    |  54%    4 MB
|============================================                                   |  55%    4 MB
|============================================                                   |  56%    4 MB
|=============================================                                  |  56%    4 MB
|==============================================                                 |  57%    4 MB
|==============================================                                 |  58%    4 MB
|===============================================                                |  58%    4 MB
|===============================================                                |  59%    4 MB
|================================================                               |  60%    4 MB
|================================================                               |  61%    4 MB
|=================================================                              |  61%    4 MB
|==================================================                             |  62%    4 MB
|==================================================                             |  63%    4 MB
|===================================================                            |  63%    4 MB
|===================================================                            |  64%    4 MB
|====================================================                           |  65%    4 MB
|====================================================                           |  66%    4 MB
|=====================================================                          |  66%    5 MB
|======================================================                         |  67%    5 MB
|======================================================                         |  68%    5 MB
|=======================================================                        |  68%    5 MB
|=======================================================                        |  69%    5 MB
|========================================================                       |  70%    5 MB
|========================================================                       |  71%    5 MB
|=========================================================                      |  71%    5 MB
|==========================================================                     |  72%    5 MB
|==========================================================                     |  73%    5 MB
|===========================================================                    |  73%    5 MB
|===========================================================                    |  74%    5 MB
|============================================================                   |  75%    5 MB
|============================================================                   |  76%    5 MB
|=============================================================                  |  76%    5 MB
|=============================================================                  |  77%    5 MB
|==============================================================                 |  78%    5 MB
|===============================================================                |  78%    5 MB
|===============================================================                |  79%    6 MB
|================================================================               |  80%    6 MB
|================================================================               |  81%    6 MB
|=================================================================              |  81%    6 MB
|=================================================================              |  82%    6 MB
|==================================================================             |  83%    6 MB
|===================================================================            |  83%    6 MB
|===================================================================            |  84%    6 MB
|====================================================================           |  85%    6 MB
|====================================================================           |  86%    6 MB
|=====================================================================          |  86%    6 MB
|=====================================================================          |  87%    6 MB
|======================================================================         |  88%    6 MB
|=======================================================================        |  88%    6 MB
|=======================================================================        |  89%    6 MB
|========================================================================       |  90%    6 MB
|========================================================================       |  90%    6 MB
|=========================================================================      |  91%    6 MB
|=========================================================================      |  92%    6 MB
|==========================================================================     |  93%    7 MB
|===========================================================================    |  93%    7 MB
|===========================================================================    |  94%    7 MB
|============================================================================   |  95%    7 MB
|============================================================================   |  95%    7 MB
|=============================================================================  |  96%    7 MB
|=============================================================================  |  97%    7 MB
|============================================================================== |  98%    7 MB
|===============================================================================|  98%    7 MB
|===============================================================================|  99%    7 MB
|================================================================================| 100%    7 MB

fruits_agg <- all_fruits %>% group_by(word) %>% summarise(count = n(),
            stroke_count_mean = mean(stroke_count),
            stroke_count_med = median(stroke_count),
            drawing_time_pause_sec_mean = mean(drawing_time_pause_seconds), 
            drawing_time_pause_sec_med = median(drawing_time_pause_seconds), 
            drawing_time_draw_sec_mean = mean(drawing_time_draw_seconds), 
            drawing_time_draw_sec_med = median(drawing_time_draw_seconds)) %>%
  arrange(-drawing_time_draw_sec_mean)

all_fruits %>% 
ggplot(aes(x = drawing_time_draw_seconds, y = factor(word, levels = fruits_agg$word), fill = factor(word, levels = fruits_agg$word))) +
geom_joy(size = 0.55 ) +
  theme_joy() +
  scale_x_continuous(limits=c(1, 20), expand = c(0.01, 0)) +
  scale_y_discrete(expand = c(0.01, 0)) + 
  scale_fill_brewer(guide = FALSE, palette = "Oranges", direction = -1) + 
  labs(title = 'Fruits', x = '', y = '')

ggsave('fruits.png', width = 8, height = 12)

all_fruits %>% filter(word == "banana") %>%
  ggplot(aes(x = drawing_time_seconds,  y = (..count..)/sum(..count..))) +
  geom_histogram(binwidth = 0.1, fill = "#FFD600") + 
  scale_x_continuous(limits = c(0, 20)) + 
  scale_y_continuous(labels = scales::percent) +
  theme_joy()  + 
  labs(title = "Banana Total Drawing Time", x = "", y = "")
ggsave("banana_total.png", width = 6, height = 5)

all_fruits %>% filter(word == "banana") %>%
  ggplot(aes(x = drawing_time)) +
  geom_histogram(binwidth = 100) +
  scale_x_continuous(limits = c(0, 20000))

all_fruits %>% filter(word == "banana") %>%
  ggplot(aes(x = stroke_count)) +
  geom_histogram(binwidth = 1) 

  #scale_x_continuous(limits = c(0, 20))

all_fruits %>% filter(stroke_count < 5) %>%
  ggplot(aes(x = drawing_time)) +
  geom_histogram(binwidth = 100) +
  scale_x_continuous(limits = c(0, 20000))

all_cates %>% #filter(stroke_count < 5) %>%
  ggplot(aes(x = drawing_time_seconds,  y = (..count..)/sum(..count..))) +
  geom_histogram(binwidth = 1) +
  scale_y_continuous(labels = scales::percent) +
  scale_x_continuous(limits = c(0, 10)) + 
  theme_joy() +
  labs(title = "All Categories Total Drawing Time (binwidth = 1)", y = '', x = '')
ggsave('all_total_1.png', width = 6, height = 5)

bananas %>% filter(recognized == 'True') %>%
  ggplot(aes(x = drawing_time_draw_seconds )) +
  geom_histogram(binwidth = 0.1) +
  scale_x_continuous(limits = c(0.5, 8))

ggplot() +
  geom_histogram(data = bananas, aes(x = drawing_time_seconds, y = (..count..)/sum(..count..)), alpha = 1.0, fill = "#FFD600", binwidth = 0.1) +
  geom_histogram(data = all_fruits %>% filter(stroke_count < 5), aes(x = drawing_time_seconds, y = (..count..)/sum(..count..)), alpha = 0.6, fill = 'black', binwidth = 0.1) +
  #geom_histogram(data = dogs, aes(x = drawing_time_seconds, y = (..count..)/sum(..count..)), alpha = 0.6, fill = 'blue', binwidth = 0.1) +
  scale_x_continuous(limits = c(0, 10), breaks = seq(0,10,0.5)) + 
  scale_y_continuous(labels = scales::percent) +
  labs(title = "Banana vs All Fruit Drawing Time (binwidth = 0.1)", y = "", x = "Drawing Time (secs)") +
  theme_light()
ggsave("bananas_vs_fruit.png", width = 6, height = 5) 

mean(all_fruits$drawing_time_seconds)

[1] 6.211525

mean(dogs$drawing_time_seconds)

[1] 11.08655

getmode <- function(v) {
   uniqv <- unique(v)
   uniqv[which.max(tabulate(match(v, uniqv)))]
}
getmode(bananas$drawing_time_seconds)

[1] 2.615

summary()

bananas_two_sec %>%
  ggplot(aes(x =  factor(stroke_count))) +
  geom_bar()

bananas %>% filter(stroke_count < 4) %>%
  ggplot(aes(x = drawing_time_seconds)) +
  geom_histogram(binwidth = 0.5) +
  scale_x_continuous(limits = c(1, 8)) + 
  facet_wrap(~ stroke_count)

pears %>% filter(stroke_count < 6) %>%
  ggplot(aes(x = drawing_time_seconds)) +
  geom_histogram(binwidth = 0.5) +
  scale_x_continuous(limits = c(0, 10)) + 
  facet_wrap(~ stroke_count)

dogs %>% filter(stroke_count < 4) %>% filter(drawing_time_seconds < 20) %>% filter(recognized == 'True')  %>%
  ggplot(aes(x = drawing_time_seconds)) +
  geom_histogram(binwidth = 0.1) +
  scale_x_continuous(limits = c(4, 11)) +
  facet_wrap(~ stroke_count)

dogs_m <- dogs %>% mutate(sec = floor(drawing_time_seconds), time_in_sec = drawing_time_seconds - sec)

dogs_m %>% filter(stroke_count < 4) %>% filter(sec %in% c(4,5,6,7,8,9, 10, 11, 12, 13, 14, 15)) %>%
  ggplot(aes(x = time_in_sec)) +
  geom_histogram(binwidth = 0.01) +
  #scale_x_continuous(limits = c(1, 20)) +
  facet_wrap(~ sec)

dogs %>% filter(stroke_count < 4) %>% filter(drawing_time_seconds < 20) %>%
  ggplot(aes(x = drawing_time_seconds)) +
  geom_area()

Error in eval(expr, envir, enclos) : object 'y' not found

bananas_by_time <- bananas %>% group_by(round_time, stroke_count) %>% summarise(n = n()) %>% mutate(freq = n / sum(n))

bananas_by_time %>% filter(round_time < 20) %>% filter(stroke_count < 10) %>%
  ggplot(aes(x = round_time, fill = factor(stroke_count), y = freq)) +
  geom_bar(stat = 'identity')

animals_meta <- meta %>% filter(animal == TRUE)
all_animals <- read_names(animals_meta)

library(RColorBrewer)

animals_agg <- all_animals %>% group_by(word) %>% summarise(count = n(),
            stroke_count_mean = mean(stroke_count),
            stroke_count_med = median(stroke_count),
            drawing_time_pause_sec_mean = mean(drawing_time_pause_seconds), 
            drawing_time_pause_sec_med = median(drawing_time_pause_seconds), 
            drawing_time_draw_sec_mean = mean(drawing_time_draw_seconds), 
            drawing_time_draw_sec_med = median(drawing_time_draw_seconds)) %>%
  arrange(-drawing_time_draw_sec_mean)

all_animals %>% 
ggplot(aes(x = drawing_time_draw_seconds, y = factor(word, levels = animals_agg$word), fill = factor(word, levels = animals_agg$word))) +
geom_joy(size = 0.55 ) +
  theme_joy() +
  scale_x_continuous(limits=c(1, 18), expand = c(0.01, 0)) +
  scale_y_discrete(expand = c(0.01, 0)) + 
  #scale_fill_brewer(guide = FALSE, palette = "YlGnBu", direction = -1) + 
  scale_fill_manual(values=rep(brewer.pal(6,"GnBu"),times=9), guide = FALSE) +
  labs(title = 'Animals', x = '', y = '')
ggsave('animals.png', width = 8, height = 12)

all_cates <- read_names(meta)

all_agg <- all_cates %>% group_by(word) %>% summarise(n = n(),
            scount_me = mean(stroke_count),
            scount_md = median(stroke_count),
            d_pause_me = mean(drawing_time_pause_seconds), 
            d_pause_md = median(drawing_time_pause_seconds), 
            d_draw_me = mean(drawing_time_draw_seconds), 
            d_draw_md = median(drawing_time_draw_seconds),
            d_total_md = median(drawing_time_seconds),
            d_total_me = mean(drawing_time_seconds)) %>%
  arrange(-d_draw_me)

all_agg_rec <- all_cates %>% filter(recognized == 'True') %>% group_by(word) %>% summarise(n = n(),
            scount_me = mean(stroke_count),
            scount_md = median(stroke_count),
            d_pause_me = mean(drawing_time_pause_seconds), 
            d_pause_md = median(drawing_time_pause_seconds), 
            d_draw_me = mean(drawing_time_draw_seconds), 
            d_draw_md = median(drawing_time_draw_seconds),
            d_total_md = median(drawing_time_seconds),
            d_total_me = mean(drawing_time_seconds)) %>%
  arrange(-d_draw_me)

all_agg %>%
  ggplot(aes(x = drawing_time_draw_sec_mean, y = drawing_time_draw_sec_med)) +
  geom_point()

library(ggbeeswarm)

all_agg %>% 
  ggplot(aes(x = drawing_time_draw_sec_med, y = '')) + 
  geom_beeswarm(groupOnX = FALSE, cex=2.8, size = 2) +
  labs(x = '', y = '')
  #geom_quasirandom(groupOnX = FALSE, varwidth = TRUE)
  

write_csv(all_agg, 'all_aggregates.csv')

All Categories

data_path <- './data/'
stat_files <- list.files(path = data_path, pattern = '\\.stats\\.csv')
length(stat_files)
all_cates <- vector("list",length(stat_files))
length(all_cates)

for(i in 1:length(stat_files)) {
  stat_file <- stat_files[i]
  full_path <- paste(data_path, stat_file, sep='')
  print(full_path)
  cate_org <- read_csv(full_path)
  cate <- prepare_data(cate_org)
  all_cates[[i]] <- cate
  
}

all_cates_df <- do.call("rbind", all_cates)

all_cates_df_clean <- prepare_data(all_cates_df)

all_cates_df_clean_rec <- all_cates_df_clean %>% filter(recognized == 'True')

all_cates_mean <- all_cates_df_clean_rec %>% group_by(word) %>% 
  summarise(count = n(),
            stroke_count_mean = mean(stroke_count),
            stroke_count_med = median(stroke_count),
            drawing_time_pause_sec_mean = mean(drawing_time_pause_seconds), 
            drawing_time_pause_sec_med = median(drawing_time_pause_seconds), 
            drawing_time_draw_sec_mean = mean(drawing_time_draw_seconds), 
            drawing_time_draw_sec_med = median(drawing_time_draw_seconds)) %>%
  arrange(-drawing_time_draw_sec_mean)

all_cates_df_clean_rec$word <- factor(all_cates_df_clean_rec$word, levels = all_cates_df_clean_rec$word[order(all_cates_df_clean_rec$word)])


all_cates_df_clean_rec %>% filter(drawing_time_draw_seconds <= 20) %>%
  ggplot(aes(x = drawing_time_draw_seconds, y = word ))+
  geom_joy(scale = 4) + theme_joy()

all_cates_df %>% filter(word == 'dog') %>% count()

all_cates_df_clean_rec %>% filter(word == 'dog') %>% filter(drawing_time_draw_seconds <= 20) %>%
 ggplot(aes(x = drawing_time_draw_seconds)) +
  geom_histogram(binwidth = 1) + 
  labs(title = "Histogram of Drawing Time")